Targeting epigenetic pathways has gained momentum as a therapeutic strategy in cancer. DNA methylation inhibitors have shown substantial activity in hematologic malignancies and two of them, azacytidine and 5- aza-2'-deoxycytidine (DAC) are now FDA approved for the treatment of the myelodysplastic syndrome (MDS). Histone deacetylase inhibitors (HDACi) have shown definite though more modest activity in hematologic neoplasia as well. There is in-vitro synergy between these two epigenetic pathways, at least at the level of gene reactivation. In a randomized phase III study of DAC (the most potent DMA methylation inhibitor available), we found that methylation decreases in the treated arm while it increases in the supportive care arm, and responders had more CpG island hypomethylation than non-responders. In a subsequent randomized phase II study, we identified a dose schedule that results in a complete remission (CR) rate of about 40% in MDS (by IWG criteria). Responses were associated with P15 induction, and patients with the highest levels of P15 after treatment were most likely to achieve a CR. We recently completed a phase I study combining DAC with valproic acid (VPA), an orally available HDACi, initially approved as an anti-seizure medication but that also has demonstrated anti-neoplastic activity. The DAC/VPA combination is safe and results in clinical responses in relapsed/refractory hematologic malignancies. In-vitro data showed histone hyperacetylation, a molecular phenomenon not observed after single agent DAC as well as more pronounced activation of P15 following combination therapy. These data led us to suggest the hypothesis that combined epigenetic therapy with DAC+VPA is more effective molecularly and clinically than DAC alone in the treatment of MDS. To test this hypothesis, we propose the following specific aims: (i) A randomized phase II clinical trial of DAC+VPA compared to DAC alone in MDS, with primary clinical endpoints of efficacy (CR, survival) and toxicity, (ii) The combination of DAC+VPA will be more effective at activating silenced gene expression (as measured by a panel of genes tested by qPCR, including P15) and (iii) CpG island methylation (as measured by pyrosequencing of a panel of genes) will decrease after treatment and correlate with responses to the combination. This grant will support the first testing, in a randomized setting, of combined epigenetic therapy, with both molecular and clinical endpoints.